Microcystic Inner Nuclear Layer Abnormalities and Neuromyelitis Optica

ORIGINAL CONTRIBUTION Microcystic Inner Nuclear Layer Abnormalities and Neuromyelitis Optica

Jeffrey M. Gelfand, MD; Bruce A. Cree, MD, PhD, MRC; Rachel Nolan, BA; Sam Arnow, BS; Ari J. Green, MD, MCR

Importance: Microcystic abnormalities involving the in- function. The hypothesis was generated prior to the data ner nuclear layer of the retina occurs in a subset of pa- being reviewed and analyzed. tients with multiple sclerosis, most commonly in eyes previously affected by symptomatic optic neuritis. Acute optic Results: Microcystic changes were identified in 5 of 25 neuritis is a cardinal manifestation of neuromyelitis op- patients with NMO (20%) and 7 of 48 total eyes, includ- tica (NMO). To our knowledge, microcystic inner nuclear ing 7 of 29 eyes (24%) previously affected by optic neu- layer abnormalities have not been investigated in NMO. ritis. Microcystic changes occurred exclusively in eyes with a history of acute symptomatic optic neuritis (100% Objective: To establish whether microcystic inner of eyes with microcystic changes had experienced prior nuclear layer abnormalities occur in NMO. optic neuritis compared with 71% of NMO eyes without microcystic abnormalities). There were no significant dif- Design: Observational, retrospective study. ferences between patients with NMO with and without microcystic changes in terms of age, sex, and aquaporin Setting: University of California at San Francisco Mul- 4–IgG antibody status. The mean age in this cohort was tiple Sclerosis Center (academic specialty clinic). 44 years (range, 13-81 years); 84% were women; 80% were aquaporin 4–IgG seropositive; and the median Ex- Patients: Twenty-five consecutive patients with NMO based on 2006 diagnostic criteria or with NMO spec- panded Disability Status Scale score was 4.0 (interquar- trum disease (defined by seropositivity for anti– tile range, 3.0-6.5). aquaporin 4 IgG in the context of a single episode of transverse myelitis or optic neuritis). Conclusions and Relevance: Microcystic inner nuclear layer pathology occurs in a proportion of patients with Exposure: Spectral-domain optical coherence tomog- NMO in eyes previously affected by acute optic neuritis. raphy. Additional research is needed to understand the cause of this retinal pathology and determine whether it con- Main Outcomes and Measures: Identification of mi- tributes to persistent visual disability in patients with NMO crocystic inner nuclear layer pathology on spectral- following optic neuritis. domain optical coherence tomography. Multivariable linear regression was used to examine associations between JAMA Neurol. 2013;70(5):629-633. Published online microcystic changes and measures of retinal structure and March 25, 2013. doi:10.1001/jamaneurol.2013.1832

EUROMYELITISOPTICA Optic neuritis is a defining feature of (NMO) is an inflamma- NMO and typically leads to optic nerve de- tory disease of the cen- myelination,11 profound retinal nerve fi- tral nervous system char- ber layer (RNFL) axonal loss,12-16 retinal acterized by injury to the ganglion cell loss,17 and severe visual dis- optic nerve, spinal cord, brainstem, and ability.18-20 Optic neuropathy in NMO is N 1,2 brain parenchyma. Once thought to be also sometimes associated with peripap- an aggressive variant of multiple sclero- illary retinal vascular abnormalities, in- sis (MS), NMO is now known to have a cluding focal arteriolar narrowing and at- Author Affiliations: Multiple distinctive pathophysiology character- tenuation of the peripapillary vascular Author Aff Sclerosis Center, Department of ized, in part, by an antibody-mediated tree.15 Sclerosis Ce Neurology (Drs Gelfand, Cree, attack against central nervous system aqua- Microcystic macular edema, a novel Neurology and Green, Ms Nolan, and porin 4,3-5 a water channel expressed by retinal phenotype characterized by hon- and Green, Mr Arnow) and Department of 6-8 Mr Arnow) Ophthalmology, astrocytes. Serum IgG-1 antibodies to eycombed, microcystic abnormalities of Ophthalmo Neuro-Ophthalmology Service aquaporin 4 are detectable in about 80% the inner nuclear layer on optical coher- Neuro-Oph (Dr Green), University of of patients with NMO and are highly spe- ence tomography (OCT), was recently (Dr Green) California, San Francisco. cific for the diagnosis.3,9,10 identified and confirmed to occur in about California,

JAMA NEUROL/ VOL 70 (NO. 5), MAY 2013 WWW.JAMANEURO.COM 629

©2013 American Medical Association. All rights reserved. Downloaded From: http://archneur.jamanetwork.com/ by a University of California - San Francisco User on 11/05/2013 5% of patients with MS.21,22 Microcystic inner nuclear layer to read at 20/200). Color vision was tested using Hardy-Rand- abnormalities have also been observed histologically fol- Rittler color plates.32 lowing optic nerve transection23-25 and as rare findings in other types of human optic nerve injury,26 but this find- STATISTICAL ANALYSIS ing has not been consistently replicated in other model systems of optic nerve injury.27,28 Multivariable linear regression was used to examine associa- In this study, we sought to establish whether micro- tionsofmicrocysticinnernuclearlayerpathologywithfovealthick- cystic inner nuclear layer pathology occurs in NMO. ness, macular volume, RNFL thickness, visual acuity, and EDSS, with models adjusting for age and sex and, when appropriate, history of optic neuritis. Within-patient intereye correlations were METHODS accounted for using the robust clustered sandwich estimator. Dif- ferences in baseline variables between groups were analyzed using PARTICIPANTS the t test for age, the ␹2 test for sex, the Fisher exact test for optic neuritis, and the Wilcoxon rank-sum test for EDSS and time from We retrospectively identified consecutive patients with NMO first and most recent acute optic neuritis. Stata version 12.0 (Stata- based on 2006 diagnostic criteria29 or with NMO spectrum dis- Corp) was used for all analyses. ease (defined by seropositivity for anti–aquaporin 4 IgG in the context of a single episode of transverse myelitis or optic neu- RESULTS ritis) who had received spectral-domain (SD)–OCT imaging at our referral center as part of routine clinical care between 2008 and 2012. Exclusion criteria included diabetes mellitus, glau- We identified 23 patients with NMO and 2 patients with coma, uveitis, eye trauma, or other retinal disease. A history of acute optic neuritis was defined clinically and was obtained by NMO spectrum disorder (both with optic neuritis and participant report and medical record review. The Expanded positive anti–aquaporin 4 IgG) who had received SD- Disability Status Scale (EDSS) score,30 a measure of functional OCT imaging in our laboratory as part of routine clini- disability typically used in MS but also potentially informative cal care between 2008 and 2012. The mean (SD) age of in NMO, was assigned by the treating neurologist. Aquaporin this cohort was 44.2 (18.5) years (range, 13-81 years); 4 antibody testing was obtained from one of several commer- 84% were women; 15 (60%) were white (of whom 2 were cial laboratories. of Hispanic ethnicity), 7 (28%) were Asian, and 3 (12%) were black; 80% were aquaporin 4–IgG seropositive; and ETHICAL CONSIDERATIONS the median EDSS score was 4.0 (interquartile range, 3.0- 6.5). Two eyes were excluded from analysis owing to dif- All patients provided written informed consent. The Univer- ficulty obtaining interpretable scans in the context of se- sity of California at San Francisco Committee on Human Re- vere visual loss. search approved the study protocol. Microcystic inner nuclear layer pathology was identified on SD-OCT in 5 patients (20%) and in 7 of 48 eyes OCT (15%), including 7 of 29 eyes (24%) previously affected Spectral-domain OCT imaging was obtained using the Heidel- by optic neuritis (Table; Figure; eFigure, http://www berg Spectralis OCT system. For evaluation of the macula, we .jamaneuro.com). When compared with patients with MS used raster scans of the macula (20Њϫ15Њ) consisting of 19 line who underwent this identical diagnostic analysis at our scans in a rectangular section. A 360Њ peripapillary RNFL B- institution,21 microcystic changes were more common in scan was obtained at a distance of 3.4 mm from the center of patients with NMO (20%, 5 of 25) in our data set than the papilla and mean RNFL thickness was calculated using in those with MS (5%, 15 of 318) (P=.01; Fisher exact Heidelberg software.31 Laboratory-quality control measures in- test). Microcystic abnormalities occurred exclusively in cluded review of all source images to ensure adequate signal eyes previously affected by acute optic neuritis (100% of to noise, a good fit, and the absence of other significant retinal eyes with microcystic changes had prior optic neuritis or optic nerve pathology. Our laboratory target for macular scans compared with 71% of NMO eyes without microcystic was an automatic real-time (the number of B-scan images used to construct the final A-scan image) of 48 and a quality score changes; P=.17). The median time from the first optic (a measure of signal to noise) of 20. A trained technician ob- neuritis episode in the group with microcystic changes tained all OCTs. Microcystic inner nuclear layer abnormali- was 6.4 years (interquartile range, 5.6-10.2 years) com- ties were defined as cystic, honeycombed, lacunar areas of hy- pared with 4.5 years (interquartile range, 0.4-8.6 years) poreflectivity with clear boundaries on SD-OCT macular raster in NMO optic neuritis eyes without microcystic changes scan images, excluding speckling artifact.21 (P=.16). There were no appreciable differences in age, sex, aquaporin 4 antibody seropositivity, or EDSS score VISUAL TESTING between patients with and without microcystic changes. High-contrast and low-contrast vision and color vi- High-contrast visual acuity was measured using a computer- sion trended lower in eyes with microcystic changes, but ized Early Treatment Diabetic Retinopathy Study chart this difference did not reach statistical significance. The (ProVideo) under standard lighting conditions in the same ex- total RNFL was thinner in eyes with microcystic changes amination room using a standardized laboratory protocol. Low- contrast visual acuity was obtained using a computerized chart compared with all NMO eyes without microcystic changes (ProVideo) at 20/200, assigning a score between 0 and 100 based (18.9 ␮m less; 95% CI, Ϫ3.0 to Ϫ34.8; P = .02; adjust- on the lowest contrast at which participants could read the let- ment for age and sex did not influence the magnitude of ters (100 points for reading at 1.2% contrast, 5 points for the the effect or the statistical significance). The total RNFL ability to read at 100% contrast, and 0 points for the inability was 6.6 ␮m lower in eyes with microcystic changes com-

JAMA NEUROL/ VOL 70 (NO. 5), MAY 2013 WWW.JAMANEURO.COM 630

Neuromyelitis Optica, Median (IQR) Without Microcystic With Microcystic Inner Nuclear Changes Inner Nuclear Changes P Value Age, mean (SD), y 45.6 (19) 38.6 (17) .46 Female, % 85 80 Ͼ.99 Expanded Disability Status Scale score 4 (3.0 to 6.5) 6 (3.0 to 6.5) .81 Eyes with prior acute optic neuritis, No. (%) 29 (71) 7 (100) .17 Optic neuritis episodes per affected eye 1 (1 to 2) 2 (1 to 3) .49 Range 1-6 1-4 Time since first ever acute optic neuritis, y 4.5 (0.4 to 8.6) 6.4 (5.6 to 10.2) .16 Time since most recent acute optic neuritis, y 1.5 (0.2 to 4.0) 5.6 (0.4 to 6.1) .34 High-contrast visual acuity, LogMAR unitsa All eyes 0.1 (Ϫ0.1 to 1.6) 1.6 (0 to 3.0) .08 Eyes with prior optic neuritis 0.5 (0 to 3.0) 1.6 (0 to 3.0) .36 Low-contrast acuity 56 (0 to 86) 0 (0 to 39) .22 Color vision (using HRR plates, calculated as best out of 19) 7 (0 to 16) 0 (0 to 17) .73 Total retinal nerve fiber layer thickness, mean (SD), µm All eyes 68.0 (28.2) 49.1 (16.2) .02 Eyes with prior optic neuritis 55.7 (24.2) 49.1 (16.2) .41 Total macular volume, mean (SD), mm3 All eyes 2.83 (0.20) 2.85 (0.14) .75 Eyes with prior optic neuritis 2.76 (0.19) 2.85 (0.14) .18 Foveal thickness, mean (SD), µm All eyes 255.3 (20.2) 264.0 (15.5) .19 Eyes with prior optic neuritis 254.7 (21.8) 264.0 (15.5) .19

Abbreviations: HRR, Hardy-Rand-Rittler; IQR, interquartile range; OCT, optical coherence tomography. a The LogMAR scale is commonly used for statistical analysis of visual acuity and is calculated as the logarithm (base 10) of the decimal value on the Snellen scale. For example, 20/20 acuity is 0 on the LogMAR scale, whereas 20/200 is 1.0 on the LogMAR scale and 20/10 is Ϫ0.30.33

200 µm Unaffected healthy control

200 µm NMO with microcystic inner nuclear layer abnormalities

Figure. For comparison, scans of a healthy patient vs an affected patient. A, Representative macular scan of an unaffected healthy woman. B, Microcystic macular abnormalities of the inner nuclear layer on spectral-domain optical coherence tomography in a patient with neuromyelitis optica (NMO) in an eye with prior optic neuritis. The rectangular box highlights the area of microcystic abnormalities, and the thin arrows point to exemplary cysts. The asterisks denote the inner nuclear layer. Note that the retinal nerve fiber layer (arrowheads) is also thinner in the patient with NMO.

pared with NMO eyes with prior optic neuritis without to exclude the possibility of no association. Foveal thick- microcystic changes (95% CI, 9.7 to Ϫ23.0; P = .41). Total ness also trended higher in eyes with microcystic changes macular volume trended 0.09 mm3 higher (95% CI, 0.22 (8.7 ␮m higher compared with eyes without microcys- to Ϫ0.05; P = .18) in the group with microcystic changes tic changes; 95% CI, Ϫ4.4 to 21.7; P = .19; and 9.3 ␮m compared with NMO optic neuritis eyes without micro- higher than optic neuritis eyes without microcystic cystic changes, but the confidence intervals were too broad changes; 95% CI, Ϫ4.9 to 23.4; P = .19).

JAMA NEUROL/ VOL 70 (NO. 5), MAY 2013 WWW.JAMANEURO.COM 631

©2013 American Medical Association. All rights reserved. Downloaded From: http://archneur.jamanetwork.com/ by a University of California - San Francisco User on 11/05/2013 COMMENT patients with NMO studied were similar to other published cohorts.40-42 In conclusion, microcystic abnormalities of the in- The main finding of this study was that microcystic ab- ner nuclear layer on SD-OCT are relatively common in normalities of the inner nuclear layer were detectable on NMO eyes previously affected by optic neuritis. Further SD-OCT in 20% of patients with NMO and 24% of NMO research is needed to understand the cause of this reti- optic neuritis eyes. Furthermore, this microcystic inner nal pathology and elucidate whether it contributes to per- nuclear layer pathology in NMO was observed exclu- sistent visual disability in patients with NMO following sively in eyes previously affected by optic neuritis. optic neuritis. Microcystic inner nuclear layer abnormalities on OCT were recently observed in about 5% of patients with MS Accepted for Publication: November 16, 2012. and were associated with greater MS disease severity and Published Online: March 25, 2013. doi:10.1001 decreased visual acuity.21,22 While the sample size of the /jamaneurol.2013.1832 NMO group with microcystic abnormalities was small, Correspondence: Jeffrey M. Gelfand, MD, University of which could have influenced the statistical analysis, the California, San Francisco, Multiple Sclerosis Center, Neu- results from this study suggest that the prevalence of mi- rosciences Research Bldg, PO Box 3206, San Francisco, crocystic inner nuclear layer pathology on SD-OCT ap- CA 94158 ([email protected]). pears to be substantially higher in NMO. This may re- Author Contributions: Study concept and design: Gelfand, late to differences in pathophysiology between the 2 Cree, Nolan, and Green. Acquisition of data: Gelfand, Cree, conditions, as optic neuritis in NMO tends to be more Nolan, and Green. Analysis and interpretation of data: aggressive and more destructive, as well as leads to greater Gelfand, Cree, Arnow, and Green. Drafting of the manu- visual disability than it does in MS.13,14,18,20 As microcys- script: Gelfand. Critical revision of the manuscript for im- tic inner nuclear layer abnormalities were observed ex- portant intellectual content: Cree, Nolan, Arnow, and clusively in NMO eyes previously affected by optic neu- Green. Statistical analysis: Gelfand. Obtained funding: ritis, these results support the idea that optic nerve injury Green. Administrative, technical, and material support:Cree, is necessary (but not necessarily sufficient) to cause mi- Nolan, and Green. Study supervision: Cree and Green. crocystic inner nuclear layer pathology. Conflict of Interest Disclosures: Dr Gelfand has received The cause of microcystic inner nuclear layer pathol- honoraria from the National Multiple Sclerosis Society for ogy in MS and NMO is unknown, but it may relate to patient education. Dr Cree has received personal compen- transsynaptic degeneration following optic nerve in- sation for consulting with Biogen-Idec, EMD Serono, Novar- jury,23,24 local disruptions in the blood-retinal barrier, fo- tis, Sanofi-Aventis, and Teva Neurosciences. Dr Green has cal retinal inflammation, and/or microglial activation in provided consulting services for Prana Pharmaceuticals, response to neuronal and axonal loss and/or injury. Some Novartis, Biogen, Roche, and Acorda Pharmaceuticals. Dr have proposed calling this OCT phenotype microcystic Green has served on an end point adjudication committee macular degeneration of optic neuropathy.34 Interest- for a Biogen-sponsored trial and provided expert legal ad- ingly, in the context of NMO, aquaporin 4 water chan- vice for Mylan Pharmaceuticals. nels are abundantly expressed by Mu¨ller cells, retinal glia Funding/Support: This study was supported by fund- located within and around the inner nuclear layer that ing from the American Academy of Neurology Clinical are involved in blood-retinal barrier functioning35 and os- Research Training Fellowship (Dr Gelfand) and grant KL2 motic regulation.8 Aquaporin 4 channels also appear to RR 024130 from the National Institutes of Health (Dr serve a neuroprotective function following acute ische- Green). mic injury in mice.36 More research is needed to deter- Online-Only Material: The eFigure is available at http: mine whether Mu¨ller cells are ever an antigenic target //www.jamaneuro.com. in NMO or whether Mu¨ller cells are involved in microcystic pathology in MS and NMO. REFERENCES This study has some important limitations. First, the sample size was relatively small, which limited statisti- 1. Cree B. Neuromyelitis optica: diagnosis, pathogenesis, and treatment. Curr Neu- cal power, especially for determining associations with rol Neurosci Rep. 2008;8(5):427-433. other clinical variables (including visual function and 2. Jacob A, Matiello M, Wingerchuk DM, Lucchinetti CF, Pittock SJ, Weinshenker overall disease disability). Second, comprehensive oph- BG. Neuromyelitis optica: changing concepts. J Neuroimmunol. 2007;187(1-2): 126-138. thalmological examinations were not routinely per- 3. Lennon VA, Wingerchuk DM, Kryzer TJ, et al. A serum autoantibody marker of formed. Third, the profound degree of visual disability neuromyelitis optica: distinction from multiple sclerosis. Lancet. 2004;364 in NMO may have made it more difficult to discern po- (9451):2106-2112. tential contributions of this retinal pathology to persis- 4. Roemer SF, Parisi JE, Lennon VA, et al. Pattern-specific loss of aquaporin-4 im- tent visual dysfunction. And finally, the possible effects munoreactivity distinguishes neuromyelitis optica from multiple sclerosis.Brain. 2007;130(pt 5):1194-1205. of disease-modifying therapy on microcystic changes were 5. Bradl M, Misu T, Takahashi T, et al. Neuromyelitis optica: pathogenicity of pa- not assessed. As patients with NMO are typically, but not tient immunoglobulin in vivo. Ann Neurol. 2009;66(5):630-643. always,37-39 treated with different disease-modifying 6. Papadopoulos MC, Verkman AS. Aquaporin 4 and neuromyelitis optica. Lancet therapies than patients with MS, we cannot exclude a Neurol. 2012;11(6):535-544. 7. Li J, Patil RV, Verkman AS. Mildly abnormal retinal function in transgenic mice contribution of therapy to the observed difference in without Mu¨ller cell aquaporin-4 water channels. Invest Ophthalmol Vis Sci. 2002; microcystic changes between MS and NMO. A relative 43(2):573-579. strength of this study is that the demographics of the 8. Pannicke T, Wurm A, Iandiev I, et al. Deletion of aquaporin-4 renders retinal glial

JAMA NEUROL/ VOL 70 (NO. 5), MAY 2013 WWW.JAMANEURO.COM 632

©2013 American Medical Association. All rights reserved. Downloaded From: http://archneur.jamanetwork.com/ by a University of California - San Francisco User on 11/05/2013 cells more susceptible to osmotic stress. J Neurosci Res. 2010;88(13):2877- 25. Gills JPJ, Wadsworth JAC. Retrograde transsynaptic degeneration of the inner 2888. nuclear layer of the retina. Invest Ophthalmol Vis Sci. 1967;6(4):437-448. 9. McKeon A, Fryer JP, Apiwattanakul M, et al. Diagnosis of neuromyelitis spec- 26. Balk LJ, Killestein J, Polman CH, Uitdehaag BMJ, Petzold A. Microcystic macu- trum disorders: comparative sensitivities and specificities of immunohistochemi- lar oedema confirmed, but not specific for multiple sclerosis. Brain. 2012;135 cal and immunoprecipitation assays. Arch Neurol. 2009;66(9):1134-1138. (pt 12):e226-author reply e227. 10. Magan˜a SM, Pittock SJ, Lennon VA, Keegan BM, Weinshenker BG, Lucchinetti 27. Darby JE, Carr RA, Beazley LD. Retinal ganglion cell death during regeneration CF. Neuromyelitis optica IgG serostatus in fulminant central nervous system in- of the frog optic nerve is not accompanied by appreciable cell loss from the in- flammatory demyelinating disease. Arch Neurol. 2009;66(8):964-966. ner nuclear layer. Anat Embryol (Berl). 1990;182(5):487-492. 11. Mandler RN, Davis LE, Jeffery DR, Kornfeld M. Devic’s neuromyelitis optica: a 28. Koma´romy AM, Brooks DE, Ka¨llberg ME, et al. Long-term effect of retinal gan- clinicopathological study of 8 patients. Ann Neurol. 1993;34(2):162-168. glion cell axotomy on the histomorphometry of other cells in the porcine retina. 12. de Seze J, Blanc F, Jeanjean L, et al. Optical coherence tomography in neuro- J Glaucoma. 2003;12(4):307-315. myelitis optica. Arch Neurol. 2008;65(7):920-923. 29. Wingerchuk DM, Lennon VA, Pittock SJ, Lucchinetti CF, Weinshenker BG. Re- 13. Ratchford JN, Quigg ME, Conger A, et al. Optical coherence tomography helps vised diagnostic criteria for neuromyelitis optica. Neurology. 2006;66(10): differentiate neuromyelitis optica and MS optic neuropathies. Neurology. 2009; 1485-1489. 73(4):302-308. 30. Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an Expanded Dis- 14. Naismith RT, Tutlam NT, Xu J, et al. Optical coherence tomography differs in ability Status Scale (EDSS). Neurology. 1983;33(11):1444-1452. neuromyelitis optica compared with multiple sclerosis. Neurology. 2009;72 31. Gelfand JM, Goodin DS, Boscardin WJ, Nolan R, Cuneo A, Green AJ. Retinal axo- (12):1077-1082. nal loss begins early in the course of multiple sclerosis and is similar between 15. Green AJ, Cree BA. Distinctive retinal nerve fibre layer and vascular changes in progressive phenotypes. PLoS One. 2012;7(5):e36847. neuromyelitis optica following optic neuritis.J Neurol Neurosurg Psychiatry. 2009; 32. Villoslada P, Cuneo A, Gelfand J, Hauser SL, Green A. Color vision is strongly 80(9):1002-1005. associated with retinal thinning in multiple sclerosis. Mult Scler. 2012;18(7): 16. Monteiro MLR, Fernandes DB, Apo´stolos-Pereira SL, Callegaro D. Quantifica- 991-999. tion of retinal neural loss in patients with neuromyelitis optica and multiple scle- 33. Gregori NZ, Feuer W, Rosenfeld PJ. Novel method for analyzing Snellen visual rosis with or without optic neuritis using Fourier-domain optical coherence acuity measurements. Retina. 2010;30(7):1046-1050. tomography. Invest Ophthalmol Vis Sci. 2012;53(7):3959-3966. 34. Abegg M, Zinkernagel M, Wolf S. Microcystic macular degeneration from optic 17. Syc SB, Saidha S, Newsome SD, et al. Optical coherence tomography segmen- neuropathy. Brain. 2012;135(pt 12):e225. tation reveals ganglion cell layer pathology after optic neuritis. Brain. 2012; 35. Reichenbach A, Wurm A, Pannicke T, Iandiev I, Wiedemann P, Bringmann A. 135(pt 2):521-533. Mu¨ller cells as players in retinal degeneration and edema. Graefes Arch Clin Exp 18. Papais-Alvarenga RM, Carellos SC, Alvarenga MP, Holander C, Bichara RP, Thuler LC. Clinical course of optic neuritis in patients with relapsing neuromyelitis optica. Ophthalmol. 2007;245(5):627-636. Arch Ophthalmol. 2008;126(1):12-16. 36. Da T, Verkman AS. Aquaporin-4 gene disruption in mice protects against im- 19. Merle H, Olindo S, Bonnan M, et al. Natural history of the visual impairment of paired retinal function and cell death after ischemia. Invest Ophthalmol Vis Sci. relapsing neuromyelitis optica. Ophthalmology. 2007;114(4):810-815. 2004;45(12):4477-4483. 20. Fernandes DB, Ramos RdeI, Falcochio C, Apo´stolos-Pereira S, Callegaro D, Mon- 37. Papeix C, Vidal JS, de Seze J, et al. Immunosuppressive therapy is more effec- teiro ML. Comparison of visual acuity and automated perimetry findings in pa- tive than interferon in neuromyelitis optica. Mult Scler. 2007;13(2):256-259. tients with neuromyelitis optica or multiple sclerosis after single or multiple at- 38. Palace J, Leite MI, Nairne A, Vincent A. Interferon beta treatment in neuromy- tacks of optic neuritis. J Neuroophthalmol. 2012;32(2):102-106. elitis optica: increase in relapses and aquaporin 4 antibody titers. Arch Neurol. 21. Gelfand JM, Nolan R, Schwartz DM, Graves J, Green AJ. Microcystic macular 2010;67(8):1016-1017. oedema in multiple sclerosis is associated with disease severity. Brain. 2012; 39. Kleiter I, Hellwig K, Berthele A, et al; Neuromyelitis Optica Study Group. Failure 135(pt 6):1786-1793. of natalizumab to prevent relapses in neuromyelitis optica. Arch Neurol. 2012; 22. Saidha S, Sotirchos ES, Ibrahim MA, et al. Microcystic macular oedema, thick- 69(2):239-245. ness of the inner nuclear layer of the retina, and disease characteristics in mul- 40. Wingerchuk DM, Hogancamp WF, O’Brien PC, Weinshenker BG. The clinical course tiple sclerosis: a retrospective study [published correction appears inLancet Neu- of neuromyelitis optica (Devic’s syndrome). Neurology. 1999;53(5):1107-1114. rol. 2012;11(12):1021]. Lancet Neurol. 2012;11(11):963-972. 41. Collongues N, Marignier R, Ze´phir H, et al. Long-term follow-up of neuromyeli- 23. Vanburen JM. Trans-synaptic retrograde degeneration in the visual system of tis optica with a pediatric onset. Neurology. 2010;75(12):1084-1088. primates. J Neurol Neurosurg Psychiatry. 1963;26:402-409. 42. Jarius S, Ruprecht K, Wildemann B, et al. Contrasting disease patterns in sero- 24. Gills JPJ Jr, Wadsworth JA. Degeneration of the inner nuclear layer of the retina positive and seronegative neuromyelitis optica: a multicentre study of 175 patients. following lesions of the optic nerve. Trans Am Ophthalmol Soc. 1966;64:66-88. J Neuroinflammation. 2012;9:14.

JAMA NEUROL/ VOL 70 (NO. 5), MAY 2013 WWW.JAMANEURO.COM 633